Large Eddy Simulation of Flow around Circular Cylinders on Structured and Unstructured Grids, II

Direct numerical simulation of oblique vortex shedding from a cylinder in shear flow

“…For this type of flow, where the cylinder axis is perpendicular to the upstream velocity and vorticity vectors at the same time, interesting results was reported in previous experimental (Maull and Young, 1973;Stansby, 1976;Tavoularis et al, 1987;Woo et al, 1989), numerical (Fr€ ohlich, 2000;Fr€ ohlich et al, 2001;Mukhopadhyay et al, 2002) and theoretical (Noack et al, 1991) studies. The complex nature of this flow geometry was clearly shown through the identification of numerous physical processes related to a specific vortex dynamics where oblique vortex shedding, dislocations and cell formations are the main ingredients.…”

“…The critical value of this length corresponds to the size of a cell where, by definition, the vortex shedding frequency should be constant. The formation of cells in the vortex shedding behind a body in a shear flow was already noted in previous studies (Maull and Young, 1973;Stansby, 1976;Tavoularis et al, 1987, Woo et al, 1989Fr€ ohlich, 2000;Fr€ ohlich et al, 2001;Mukhopadhyay et al, 2002). A simple theoretical model was proposed by Noack et al (1991) for this phenomenon at low Reynolds number (100 < Re < 160).…”

Development of wall models for LES of separated flows using statistical evaluations

“…Despite its simplicity, definition (18) can be motivated physically. Utyuzhnikov [44] applied the same assumption for a RANS wall model and a smoothing function to avoid a jump in the artificial viscosity l t .…”

“…This idea of Simpson [30] was developed further by Frö hlich et al [31]. Taking theoretical and empirical arguments into account, a relation was derived which was used by Frö hlich et al [17,18] in order to calculate the time-averaged wall shear stress hjs w ji on the back side of a circular cylinder. This model is valid only for the backflow region and could be solely applied if the time-averaged velocity U N is known.…”

“…It will be shown theoretically later that this wall model can predict the wall shear stress reliably only if the wall-nearest point is located in the viscous sublayer. Frö hlich et al [17,18] proposed an extension of Schumann's wall model [4] applying the integral method of Thwaites [19,20] based on empirical relations to the time-averaged values in order to obtain the time-averaged wall shear stress hjs w ji on the front part of a circular cylinder. This method is similar to the TLM of Balaras and Benocci [9] and has the important advantage that no auxiliary grid is necessary.…”